图像分割技术在中子照相中的应用研究
|
摘要
中子照相技术是无损检测技术之一,它具有其他无损检测技术无可替代的特点和优点,能够获得很多其它传统技术不能得到的重要信息。在无损检测技术中,由于中子照相具有特殊的功能,其应用范围日益扩大。在欧美和亚洲一些工业发达国家,中子照相已走上了商业化的轨道,它与X射线和γ射线照相相互补充,在工业、国防、航天、考古和生物医学等方面得到了广泛的应用,发挥了重要的作用。
中子照相图像是指在中子辐照场里,中子穿透被测物质后通过中子成像系统所获取到的图像。它携带着被测物质的内部结构等信息。由于在中子成像的过程中,不可避免地会受到γ射线污染、中子散射、CCD电子噪声等因素的影响,使得所获取到的中子照相图像会发生降质,如噪声严重、细节模糊等,不利于分析、判别被测物质的内部信息。
本论文研究工作致力于研究适合于中子照相图像的去噪声、去模糊等图像处理技术,以便能够获取关于物质内部的更多的信息,丰富中子成像无损检测技术的研究内容。为此,论文在理解中子成像原理的基础上,详细分析了引起中子照相图像降质的主要因素。
为提取中子照相图像的有用信息,便于进行图像分析研究,引入图像分割技术。图像分割是由图像处理过渡到图像分析的关键步骤,在图像工程中占据重要的位置。一方面,它是目标表达的基础,对特征测量有重要的影响。另一方面,因为图像分割及其基于分割的目标表达、特征提取和参数测量等将原始图像转化为更抽象更紧凑的形式,使得更高层的图像分析和理解成为可能。
在图像分割的基础上,围绕中子照相图像所含有的各种噪声和性质,我们借鉴X光图像处理的某些经验,采用滤波,反色,二值化,数学形态学等多种图像处理算法,并在原有算法的基础上提出降噪形态学算子,降噪效果显著。本文还针对大理石下花的中子照相图像障碍物边缘去除问题,提出了扩展的Roberts算子,并对其加以改进与Hough变换结合,得到令人满意的花的图像。
Neutron photography is one of the detecting technology without destroyed. There are some characters and advantages that others don’t have, and it can get some important messages that other traditional technology can’t get. The application of the neutron photography is enlarging in the detecting technology without destroyed because its peculiar function. In many industry countries of occident and Asia, the neutron photography is on its commercial track. It’s complementary with radiography andγ-ray imaging makes it is applied extensively in the field of industry, national defence , spaceflight, archaeology and biomedicine and have an important effect in the fields.
The image of neutron photography is formed by the imaging system when neutrons in the radiancy field traverse the detected matter. They carry information of the inner structure. In the process of the creation of image, the image is occasionally different from primitive scenery, because of the influence of various factors. It is not propitious to analyse and distinguish the information of the inner structure because of the severity yawp and illegible detail.
This paper is mainly about image processing about the smear and blurred neutron photography image, in order to get more inner information of pictures. The paper therefore analyses the main factors that depress the image qualities based on apprehending the neutron photography theory.
In the interest of getting useful information and making it easy to analyse of the neutron photography image, we adhibit image segmentation technology. It is the linchpin from image processing to image analysis and take an important degree in the image engineering. On one hand, it is the base of goal expression and has an important effect on character measure. On the other hand, image segmentation technology transform the original image into abstract form which make it possible to analyse and understand the picture intensively.
Neutron photography image have many kinds of smear and property, so we use some experience of X-ray image processing for reference and adopt filter, color reverse, mathematics morphologic processing measure to reduce smear of the image. The effect is obvious.
This paper is in allusion to wipe off the fringe in neutron photography image of the flower below marble and put forward patulous Roberts operator. Then improve on it and combine with Hough transform and get satisfying image of flower.
中子照相图像是指在中子辐照场里,中子穿透被测物质后通过中子成像系统所获取到的图像。它携带着被测物质的内部结构等信息。由于在中子成像的过程中,不可避免地会受到γ射线污染、中子散射、CCD电子噪声等因素的影响,使得所获取到的中子照相图像会发生降质,如噪声严重、细节模糊等,不利于分析、判别被测物质的内部信息。
本论文研究工作致力于研究适合于中子照相图像的去噪声、去模糊等图像处理技术,以便能够获取关于物质内部的更多的信息,丰富中子成像无损检测技术的研究内容。为此,论文在理解中子成像原理的基础上,详细分析了引起中子照相图像降质的主要因素。
为提取中子照相图像的有用信息,便于进行图像分析研究,引入图像分割技术。图像分割是由图像处理过渡到图像分析的关键步骤,在图像工程中占据重要的位置。一方面,它是目标表达的基础,对特征测量有重要的影响。另一方面,因为图像分割及其基于分割的目标表达、特征提取和参数测量等将原始图像转化为更抽象更紧凑的形式,使得更高层的图像分析和理解成为可能。
在图像分割的基础上,围绕中子照相图像所含有的各种噪声和性质,我们借鉴X光图像处理的某些经验,采用滤波,反色,二值化,数学形态学等多种图像处理算法,并在原有算法的基础上提出降噪形态学算子,降噪效果显著。本文还针对大理石下花的中子照相图像障碍物边缘去除问题,提出了扩展的Roberts算子,并对其加以改进与Hough变换结合,得到令人满意的花的图像。
Neutron photography is one of the detecting technology without destroyed. There are some characters and advantages that others don’t have, and it can get some important messages that other traditional technology can’t get. The application of the neutron photography is enlarging in the detecting technology without destroyed because its peculiar function. In many industry countries of occident and Asia, the neutron photography is on its commercial track. It’s complementary with radiography andγ-ray imaging makes it is applied extensively in the field of industry, national defence , spaceflight, archaeology and biomedicine and have an important effect in the fields.
The image of neutron photography is formed by the imaging system when neutrons in the radiancy field traverse the detected matter. They carry information of the inner structure. In the process of the creation of image, the image is occasionally different from primitive scenery, because of the influence of various factors. It is not propitious to analyse and distinguish the information of the inner structure because of the severity yawp and illegible detail.
This paper is mainly about image processing about the smear and blurred neutron photography image, in order to get more inner information of pictures. The paper therefore analyses the main factors that depress the image qualities based on apprehending the neutron photography theory.
In the interest of getting useful information and making it easy to analyse of the neutron photography image, we adhibit image segmentation technology. It is the linchpin from image processing to image analysis and take an important degree in the image engineering. On one hand, it is the base of goal expression and has an important effect on character measure. On the other hand, image segmentation technology transform the original image into abstract form which make it possible to analyse and understand the picture intensively.
Neutron photography image have many kinds of smear and property, so we use some experience of X-ray image processing for reference and adopt filter, color reverse, mathematics morphologic processing measure to reduce smear of the image. The effect is obvious.
This paper is in allusion to wipe off the fringe in neutron photography image of the flower below marble and put forward patulous Roberts operator. Then improve on it and combine with Hough transform and get satisfying image of flower.
引文
[1]魏宝杰,钟海明.中子管及其应用技术[M].长春:东北师范大学出版社,1997.86-105.
[2]裴宇阳,唐国有,郭之虞.中子照相技术及应用[J].新技术应用.2004,(5):17-22.
[3]貊大卫,刘以思,金光宇等.中子照相[M].北京:原子能出版社,1996.
[4]唐彬.影响中子射线照相性能的主要因素[J].核电子学与探测技术.2004,24(4):387-390.
[5]唐彬.图像处理技术在中子照相中的应用研究[J].核动力工程.2002,23(4):108-111.
[6]K.R.Castleman.数字图像处理[M].朱志刚,石定机.北京:电子工业出版社,1998.
[7]Rafael C.Gonzalez,Richard E.Woods.数字图像处理[M].2.阮秋琦,阮宇智.北京:电子工业出版社,2003.
[8]谢仁龙.卫星遥感图像去薄云薄雾处理技术[D].南京:南京理工大学,2004.
[9]孙光灵.基于模糊阈值的图像分割方法研究[D].合肥:合肥工业大学,2005.
[10]杨淑莹.VC++图像处理程序设计[M].北京:清华大学出版社,北京交通大学出版社,2003.
[11]求是科技.Visual C++数字图像处理典型算法及实现[M]北京:人民邮电出版社,2006.
[12]顾静良.低对比度弱小目标检测算法研究[D].北京:中国工程物理研究院,2005.
[13]夏良止.数字图像处理[M].南京:东南大学出版社,1999.
[14]田捷.实用图像分析与处理技术[M].北京:电子工业出版社,1994.
[15]何斌.VC++数字图像处理[M].北京:人民邮电出版社,2001.
[16]周长发.Visual C++图像编程[M].北京:电子工业出版社,2000.
[17]孙仲康,沈扳康.数字图像处理及应用[M].北京:国际工业出版社, 1983.
[18]R.M.Haralick and L.G.Shapiro.Survey: Image Segmentation[J].Comput.Vision,Graphics, Image Proc. 1985,29:100-132.
[19]A.Rosenfield. Connectivity in Digital Pictures[J].Gournal of the ACM, 1970,17:146-160.
[20]R.J.Wall.The Gray Level Histogram for Threshold Boundary Determination in Image Proces- Sing to the Scene Segmentation Problem in Human Chromosome Analysis[M].Ph.D.Thesis,Univer- sity of California at Los Angeles,1974.
[21]R.J.Wall,A.Klinger,and K.R.Castleman.Analysis of Image Histograms[J]. Copenhagen: Proc.2nd.Joint Conf.on Pattern Recognition(IEEE Pub.74CH-0885-4C) 1974,341-344.
[22]张永化.基于灰值腐蚀-膨胀形态学和中值二值化的车牌定位及硬件实现[D].浙江:浙江大学,2004.
[23]戴培山.其于小波变换的乳腺 X 线目标分割[D].成都:四川大学,2004.
[24]崔屹.数字图像处理技术与应用[M].北京:电子工业出版社,1997.
[25]崔屹.数学形态学方法及应用[M].北京:科学出版社,2003.
[26]刘循,游志胜.多尺度形态学图像边缘检测方法[J].光电工程.2003,3(30):25-30.
[27]杨述斌 ,彭复员 .数学 形 态学在图像处理中的应用与发展[J].武汉化工学院学报.2004,26(1):70-73.
[28]陈强,朱立新,夏德深.结合 Canny 算子的图像二值化[N].计算机辅助设计与图形学学报.2005.
[29]陈洪波.Hough 变换及改进算法与线段检测[D].广西:广西师范大学,2004.
[30]袁懿弘,吴锡生.基于去噪阈值的图像平滑模糊算法方法研究[J].计算机工程与技术.2005,2(3):45-51.
[31]李炜,黄心汉,王敏.一种快速 Hough 变换改进算法及其在图像边缘检测中的应用[J].信号处理.1999,25(2):182-187.
[32]D.P.Panda and A.Rosenfeld.Image Degmentation by Pixel CClassification in(Gray level, Edge value)Space[J].IEEE Trans.Comput.1978,C-27:875-879.
[33]R.A.Kirsh.Computer Determination of the Constituent Structure of Biological Images[J]. Computers in Biomedical Research.1971,4,315-328.
[34]W.K.Pratt and I.Kabir.Morphological Binary Image Processing with a Local Neighborhood Pipeline Processor[M].Tokyo:Computer Graphics.1984.
[35]W.K.Pratt.Digital Image Processing[M].2d ed. New York:John Wiley&Sons,1991.
[36]H.Blum.A Transformation for Extracting New Descriptors of Shape[R]. MIT Press, Cambridge,MA:Symposium Models for the Perception of Speech and Visual Form.W.Wathen- Dunn,ed.1967.
[37]J.C.Mott-Smith.Medial Axis Transforms[M].Academic Press,New York:Picture Processing and Psychopictorics,B.S.Lipkin and A.Rosenfeld,ed.1970.
[2]裴宇阳,唐国有,郭之虞.中子照相技术及应用[J].新技术应用.2004,(5):17-22.
[3]貊大卫,刘以思,金光宇等.中子照相[M].北京:原子能出版社,1996.
[4]唐彬.影响中子射线照相性能的主要因素[J].核电子学与探测技术.2004,24(4):387-390.
[5]唐彬.图像处理技术在中子照相中的应用研究[J].核动力工程.2002,23(4):108-111.
[6]K.R.Castleman.数字图像处理[M].朱志刚,石定机.北京:电子工业出版社,1998.
[7]Rafael C.Gonzalez,Richard E.Woods.数字图像处理[M].2.阮秋琦,阮宇智.北京:电子工业出版社,2003.
[8]谢仁龙.卫星遥感图像去薄云薄雾处理技术[D].南京:南京理工大学,2004.
[9]孙光灵.基于模糊阈值的图像分割方法研究[D].合肥:合肥工业大学,2005.
[10]杨淑莹.VC++图像处理程序设计[M].北京:清华大学出版社,北京交通大学出版社,2003.
[11]求是科技.Visual C++数字图像处理典型算法及实现[M]北京:人民邮电出版社,2006.
[12]顾静良.低对比度弱小目标检测算法研究[D].北京:中国工程物理研究院,2005.
[13]夏良止.数字图像处理[M].南京:东南大学出版社,1999.
[14]田捷.实用图像分析与处理技术[M].北京:电子工业出版社,1994.
[15]何斌.VC++数字图像处理[M].北京:人民邮电出版社,2001.
[16]周长发.Visual C++图像编程[M].北京:电子工业出版社,2000.
[17]孙仲康,沈扳康.数字图像处理及应用[M].北京:国际工业出版社, 1983.
[18]R.M.Haralick and L.G.Shapiro.Survey: Image Segmentation[J].Comput.Vision,Graphics, Image Proc. 1985,29:100-132.
[19]A.Rosenfield. Connectivity in Digital Pictures[J].Gournal of the ACM, 1970,17:146-160.
[20]R.J.Wall.The Gray Level Histogram for Threshold Boundary Determination in Image Proces- Sing to the Scene Segmentation Problem in Human Chromosome Analysis[M].Ph.D.Thesis,Univer- sity of California at Los Angeles,1974.
[21]R.J.Wall,A.Klinger,and K.R.Castleman.Analysis of Image Histograms[J]. Copenhagen: Proc.2nd.Joint Conf.on Pattern Recognition(IEEE Pub.74CH-0885-4C) 1974,341-344.
[22]张永化.基于灰值腐蚀-膨胀形态学和中值二值化的车牌定位及硬件实现[D].浙江:浙江大学,2004.
[23]戴培山.其于小波变换的乳腺 X 线目标分割[D].成都:四川大学,2004.
[24]崔屹.数字图像处理技术与应用[M].北京:电子工业出版社,1997.
[25]崔屹.数学形态学方法及应用[M].北京:科学出版社,2003.
[26]刘循,游志胜.多尺度形态学图像边缘检测方法[J].光电工程.2003,3(30):25-30.
[27]杨述斌 ,彭复员 .数学 形 态学在图像处理中的应用与发展[J].武汉化工学院学报.2004,26(1):70-73.
[28]陈强,朱立新,夏德深.结合 Canny 算子的图像二值化[N].计算机辅助设计与图形学学报.2005.
[29]陈洪波.Hough 变换及改进算法与线段检测[D].广西:广西师范大学,2004.
[30]袁懿弘,吴锡生.基于去噪阈值的图像平滑模糊算法方法研究[J].计算机工程与技术.2005,2(3):45-51.
[31]李炜,黄心汉,王敏.一种快速 Hough 变换改进算法及其在图像边缘检测中的应用[J].信号处理.1999,25(2):182-187.
[32]D.P.Panda and A.Rosenfeld.Image Degmentation by Pixel CClassification in(Gray level, Edge value)Space[J].IEEE Trans.Comput.1978,C-27:875-879.
[33]R.A.Kirsh.Computer Determination of the Constituent Structure of Biological Images[J]. Computers in Biomedical Research.1971,4,315-328.
[34]W.K.Pratt and I.Kabir.Morphological Binary Image Processing with a Local Neighborhood Pipeline Processor[M].Tokyo:Computer Graphics.1984.
[35]W.K.Pratt.Digital Image Processing[M].2d ed. New York:John Wiley&Sons,1991.
[36]H.Blum.A Transformation for Extracting New Descriptors of Shape[R]. MIT Press, Cambridge,MA:Symposium Models for the Perception of Speech and Visual Form.W.Wathen- Dunn,ed.1967.
[37]J.C.Mott-Smith.Medial Axis Transforms[M].Academic Press,New York:Picture Processing and Psychopictorics,B.S.Lipkin and A.Rosenfeld,ed.1970.